Managed Care Cast Presents: Advancements in Automated Insulin Delivery, Part 1

This podcast is brought to you by Insulet.

Today, we are bringing you part 1 of a 2-part sponsored podcast series discussing insulin delivery for patients with diabetes. In part 1, Mary Caffrey, associate editorial director of The American Journal of Managed Care^®, speaks to Greg Forlenza, MD, a pediatric endocrinologist at the Barbara Davis Center at the University of Colorado, about the evolution of diabetes care technology and standards for automated insulin delivery.

Listen above or through one of these podcast services:

iTunes
TuneIn
Stitcher
Spotify

TRANSCRIPT:

CAFFREY: Welcome to the AJMC® podcast Advancements in Automated Insulin Delivery. I’m Mary Caffrey, associate editorial director for The American Journal of Managed Care®. In this first of 2 segments, we will discuss the evolution of diabetes care technology and standards for automated insulin delivery, as well as important new developments that we will see in 2021, including Omnipod 5 [OP5].

Joining me is Greg Forlenza, [MD], a pediatric endocrinologist at the Barbara Davis Center [for Diabetes] at the University of Colorado [Anschutz Medical Campus]. In this segment, Dr Forlenza will discuss the prepivotal data. Today’s podcast is brought to you by Insulet. Welcome, Dr Forlenza.

FORLENZA: Hello, it’s great to be here.

CAFFREY: Dr Forlenza, can you tell our listeners about your background in diabetes care and your current research interests?

FORLENZA: Yes, I do work with children with type 1 diabetes, and my research is based on the role of technology to improve care for children and young adults with type 1 diabetes. The subject I’ve been interested in since the start of my training is the role of hybrid closed loop and automation to make care easier, and so I’ve been fortunate to get to work on the Medtronic 670G and 780G pivotal trials, the Tandem Basal-IQ and Control-IQ pivotal trials, the Insulet OP5 developmental and pivotal trials, and even the Beta Bionics pivotal trials, as well as some automated decisions support systems and a variety of other systems. It’s a really exciting time for the field and a very cool time for our patients.

CAFFREY: Well, thank you. The past few years have brought major advances in automated insulin delivery. Even consumers who don’t know much about diabetes have probably seen advertisements for continuous glucose monitors [CGMs] or automated insulin pumps, so can you tell us how insulin delivery has evolved, just in the past few years?

FORLENZA: I think it has been very, very dramatic. I always tell my trainees that it’s the only thing I do, and it can sometimes be hard to keep up with, and people [who] don’t look at the literature daily oftentimes fall very far behind. We still have a lot of patients who are using injections for insulin therapy, but one of the biggest changes is that they’re now starting to use a lot more CGM and seeing significant benefits with CGM along with injections.

The whole field for those using insulin pumps has also changed dramatically, as we’ve moved in the past few years from what is called sensor-augmented pump therapy—where people are looking at their own sensor values and adjusting their insulin dosing from their pump and their sensor value—to automated delivery, where the system is actually starting to take the human out of the loop and taking care of those minute-by-minute dosing adjustments by itself. This has been a phenomenal transition, in terms of both the burden that people feel daily about their diabetes care and also the level of control that they’re achieving.

Clinically, we’re seeing people’s time in range move from the 50% range to the 60% to the 70% and even the 80% range with these systems, and that’s 3 to 4 hours a day more that people are spending time in range, and they’re doing that with less work. And so what has been really, really exciting is that we’re seeing the collection of our patients just continue to move up the scale.…Even our poorest-performing patients are doing as well as our best performing patients were 5 to 10 years ago, because the technology is making it easier for them to do so.

CAFFREY: So, tell me about the technology itself. We’re seeing different kinds of pumps also in terms of some of the features of the pumps that are taking them from, say, the tethered pump to the tubeless pump.

FORLENZA: Tethered pumps would be traditionally used for the pump that has an infusion site and a cannula and that is worn on the belt or kept in the pocket, and those pumps can be detached. But a lot of patients, especially those who are young children or very active, athletic people, don’t like the idea of something “dangling from their body,” and so they like to have the option of an on-body pump, which we call a patch pump.

Currently, the popular one, of course, is the Omnipod pump. [Patients are] able to go wherever and do whatever without worrying about something hanging on their body [and] potentially getting snagged on a doorframe or a door handle or something like that. It provides an additional layer of normalization to the use of the technology, because it’s just on your body and you don’t have to worry about the tubing. It is a great option for a lot of people and, again, decreases that burden, increases that normalcy, and decreases a little bit of the visibility, which are all very big factors in people being happy with the technology—which is a huge factor in them using it successfully.

CAFFREY: OK, so, we’re hearing more about the term AID [automated insulin delivery]. How did the American Diabetes Association [ADA] Standards of Medical Care affect AID?

FORLENZA: So, that’s a great question. So, there are a lot of different terms we’ve used for this technology. The original term was artificial pancreas systems, which was a good engineering term, but it kind of let a lot of people think that this was something that was implanted, not just a sub-Q [subcutaneous] system.

AID is automated insulin delivery, and it’s basically any system where the pump is listening to the CGM and using some form of autopilot to adjust insulin delivery. I work with a lot of the early designers of these technologies, and they came from a lot of other fields. The guy who taught me about it used to program heat-seeking missiles. Another guy who taught me about it is designing self-driving cars for Toyota, and someone else who taught me a lot about it designed the systems that keep chemical plants from exploding. So, we’re taking these ideas that are used in all other areas of our lives, and we’re applying them to diabetes. I think that’s a really important point: You rely on control systems every day to keep you safe, you just don’t realize it—and this is another case where we’re trying to bring that technology there.

In terms of helping with the ADA standards, it’s transitioning the percentage of people who are able to meet targets to a much higher percentage, and we’re looking at this in now secondary and real-world analyses of pivotal trial data and real-world implementation data. So, a lot of people are going to be familiar with the data from T1 Diabetes Exchange, which showed that with conventional therapy, even 3 to 5 years ago, only about 20% of people were meeting ADA targets for hemoglobin A1C and revised ADA targets for time in range.

With this technology, we’re generally seeing that moving up to about 50%, so that’s a massive difference. When I’m in clinic, I can feel that on a daily basis—that it’s not just, like, 1 or 2 patients a day that are actually meeting targets; half or more of my patients that I’m seeing on a daily basis are meeting targets for hemoglobin A1C and for time in range. Just like having a self-driving car makes it easier for you to drive your car safely, having a self-driving diabetes system makes it easier for you to drive your diabetes safely, and so the percentage of people that are able to do that continues to increase because the tools are making it so that it’s less work, less burden, and less effort required.…As a result of that, I very strongly believe—and we’re conducting the studies now—that we’re going to see people having [fewer and fewer] complications, which we know is the end result of having better control.

And so, if you’re doubling the percentage of people who are meeting goals for time in range, you’re effectively halving, or at least reducing by a third, the percentage of people who are going to have major complications. And that’s going to continue to get better as we get to second- and third-generation systems, and just seeing more and more people meeting goals and having less risk of long-term complications—that is a tremendous benefit for the families.

When you tell [parents] that in clinic and you say, “Your 7-year-old is meeting goals, and we don’t think that she’s ever going to have complications as a result of diabetes”—that’s what they lose sleep about at night. So, it makes a difference in terms of payments, and it makes a difference in terms of quality of life, but that’s why I love technology—because we can achieve both.

CAFFREY: So, what kind of impact does this have on the care team and for care coordination, not just on the patient and the family, but also on everyone involved in the care for the person living with diabetes?

FORLENZA: So, that’s a great question. That’s a lot of impacts, actually. We still have to spend a fair amount of time working with the systems and optimizing success, but…it changes the conversation. It changes the conversation from “How can we keep you from failing?” and that very negative visit that a lot of people with established diabetes remember having to “How can we continue to optimize your success?” And that makes patients more likely to engage with care, because they don’t feel like every time they come into clinic, they’re going to be getting that failing report card. They’re going to be getting a good report card and just be discussing about learning the next thing.

[Having the systems] perpetually in the cloud—which we’re starting to see with more and more systems, and we’re definitely going to be seeing with OP5—enables us to have much easier access to the data, so we can spend less time helping a patient tune their system. A patient [or parent] will just email us or call us and say, “My daughter has just started soccer practice for 2 weeks; let’s make sure she’s not having lows.” I can just log in to the server, I can see all the CGM data, I can see all the insulin delivery data, I can see the pump settings, and I can say, “Let’s change this basal rate here; let’s change this carb ratio there,” and the parent can incorporate the change, and I can see that they’ve incorporated it. It just makes those interactions a lot easier, a lot less time-consuming, and also a lot more fulfilling. It’s a great situation, where technology can just give us a positive and help make it better for everyone, more successful, and easier.

CAFFREY: So, what’s coming in 2021 that’s new and exciting? Can you talk about the developments in AID that are coming in the next year and what the impact will be?

FORLENZA: There are a lot of good developments that are coming in the next year. The big one that, obviously, I am very excited about is the Insulet OP5 System. It’s the system that I’ve been working on developing for 5 years, all the way from the first hospital studies we did, where we had to check finger sticks constantly, to now the at-home studies, where people have been using the system safely and successfully at home for over 6 months, some of them up to 12 months now. What’s really cool about this system is that it’s the first all-on-body system, and so the hybrid closed-loop algorithm—that self-driving car idea—actually lives on the disposable pod, and that is a big advancement. Because early on, we thought it would have to live on a cell phone, and you’d have to always have a cell phone in range.

But with this system, patients just wear the CGM, they just wear the pod, the CGM communicates directly with the pod, and the system can operate in hybrid closed-loop mode, changing insulin delivery every 5 minutes, based on their blood sugar values. [It’s] just all on your body, and the cell phone is still used to start the system, check the CGM value when you want to check, and bolus for insulin. But if you go for a jog and you leave your cell phone in your gym bag or you’re out on the court playing basketball and your cell phone is in the locker room, the system is still automating. This is the first system to be able to do that, just providing that next layer of features for people that weren’t happy with the previous offers of features, and that additional peace of mind that no matter what’s going on, the system is automating.

We’re also continuing to do work with CGMs, getting them smaller, all disposable, and lasting longer, and so that’s a really big feature also. The growth of this use of CGMs and the ease and decreased cost of the CGMs, which are going to always be the key element of these systems, is requiring that continuous sensing of data.

Nick Jonas, obviously, showing off his six-pack with a CGM during the Super Bowl, I think was a really cool moment, because it helps normalize that technology for kids and says “If cool, fit Nick Jonas can do this during the Super Bowl, I can do it too,” so I was very happy to see that.

CAFFREY: Let’s talk some more about the individual features of the Omnipod 5, such as the On-Target algorithm and HypoProtect, and the Informed Bolus Calculator. What are some of the individual features, and what can they do?

FORLENZA: Yeah, so all the systems are a little bit different, and the way that we’ve worked on designing OP5 is different from the way we worked on designing Control-IQ and 670 and 780G. That creates a little bit of a challenge for clinicians in terms of learning each system. As for OP5, the algorithm that we’re using is now called On-Target. People that have heard me speak before [might know that I used to call it Horizon, but the new name for it is On-Target, and this is an algorithm that very much favors simplicity, which I think is going to be good for most patients.

The algorithm is continually updating, adjusting basal insulin delivery, background insulin delivery, every 5 minutes in real time, and it’s continuing to adapt to the user to become a little bit more personalized, updating its personalization with every pod change—so, roughly every 3 days. And what we saw in the trials was that within about 2 or 3 pod changes—so, within about a week, a week and a half—the system dials in very, very well and just sort of smooths out background insulin delivery.

And so for a lot of very engaged patients, they’ve read stuff about doing fasting challenges where they have to not eat anything for 24 hours to figure out their exact basal insulin dosing and this kind of stuff. I have a few families in my study where they did that previously, and their No. 1 favorite thing about doing the study was that they didn’t have to do fasting challenges anymore because the system was just adapting with time, and so they were very happy with that feature. And it has been very successful, especially overnight—that’s where we really see that the system just dialing in the basal insulin delivery overnight results in people very, very infrequently needing to look at their values overnight. Patients and parents described being able to sleep overnight for the first time without waking up since diagnosis and knowing that they’re going to wake up with a normal value. That’s the other thing we hear a lot: “I just wake up and my blood sugar is normal every morning. I start each day fresh. I don’t have to chase my tail and treat a high, treat a low; no matter what was going on the previous day, every day starts out fresh.”

The Informed Bolus Calculator allows for a little bit of the adaption that some of us actually wrote about in the guidelines for Dexcom G5 when it came out, which is that the system isn’t just using cold values; it’s not just using the carbs and the current CGM value. It’s also doing a little bit of that adjustment that patients typically were trained to do in their heads, where it’s increasing or decreasing the bolus a little bit based on the CGM trend. That’s another automation feature, where people are able to now remove that layer of worry: “Do I have to account for the trend arrows myself and be my own artificial pancreas in my head?” No, the system is handling that to further attempt to improve time in range.

One of the other features about the algorithm that’s really important and sort of a change from previous designs is that we’ve enabled setting of different targets. The targets with the algorithm can be set at 10 mg/dL increments between 110, 120, 130, 140, and 150, so generally, I favor the 110 target. I think most people want impressive control, and because the system does a really good job of minimizing hypoglycemia, we’re able to be at 110 as a target and have really good control, but there are certain circumstances under which the other targets are better. For, like, toddlers overnight, young children who typically are more likely to have low blood sugars than other people, we generally run with the 150 target overnight and then the 110 target during the day. Some of my colleagues do work with senior citizens who run it at the 130 to 150 targets due to risk of lows.

I’ve had some adolescents who never remember to use the HypoProtect mode during PE [physical education], and so we would just program the higher target during their PE and kind of keep them safer there. So, there are a lot of places where having dynamic targets can be really beneficial, and I’m excited to see us have that.

And then the last feature that I’ll mention is the HypoProtect mode. This is a feature that we actually developed about midway through system development, for specific circumstances under which we want to try to minimize risk of hypoglycemia exposure, and so it does a little bit more than the idea of just raising the target. It does the raise the target to 150, but it also decreases the aggressiveness of the algorithm so that there’s essentially very, very minimal risk of hypoglycemia, provided that you turn on the HypoProtect feature beforehand.

I generally recommend it for exercise. That’s obviously the time that most people are most afraid of hypoglycemia. I have some parents and toddlers where, rather than set the target differently, they just use the HypoProtect feature overnight, depending on what the kid’s day was like. We’ve also used it for illness, where we were worried about lows, especially gastrointestinal illnesses, where we thought maybe that people weren’t absorbing carbs as well, and people have used it for other specific circumstances where they were worried about lows. And so, the collection of those features again brings us back to things that people used to have to just do in their heads, which the system is now automating or automating after 1 button push. And so it just takes that “I have to be my own pancreas; I have to think like a pancreas” to “We’ve now built that into the system, and the system does it.” So, I think that’s a really exciting development and something that has been working really well.

CAFFREY: Thank you, Dr Forlenza. We’ll stop there, and I hope our listeners can join us for the second part of our conversation about some important new prepivotal data. I also want to encourage our listeners to look for the AJMC ®Clinical Brief , which will feature Omnipod 5’s phase 2 pivotal data. The podcast and practice brief will be on ajmc.com.

For all of us at The American Journal of Managed Care®, I’m Mary Caffrey. I want to thank Insulet for bringing us today’s podcast and thank you, the listeners, for joining us.